Reinforcement Learning-Enabled Hybrid Digital–Analog Beamforming for High-Efficiency Millimeter-Wave and Terahertz Transmission in 6G Networks

The increase in demand for ultra-high-speed and low-latency communications has led to a corresponding rise in the need for more efficient and reliable transmission systems. In this work, we put forward an innovative transmission model for the millimetre-wave and terahertz bands, with the objective of enhancing the transmission efficiency and reliability of high-frequency bands in 6G communication. We put forward a two-layer architectural solution that combines digital beamforming and analog beamforming in order to effectively address the multipath effect and signal attenuation issues that arise in millimetre wave and terahertz bands. In particular, digital beamforming is responsible for the precise control of the baseband signal, with the objective of minimising interference between different users and adjusting the phase and amplitude of the signal with high precision in order to adapt to the dynamic needs of users and changes in the network environment. In contrast, analog beamforming employs an array of phase shifters within the radio frequency (RF) link to modify the physical direction of the signal, thereby reducing the associated hardware complexity and power consumption. The experimental results demonstrate that the proposed model markedly enhances data throughput, signal coverage, and energy efficiency in diverse environments.